Estimation of the catchment area of a sediment trap by means of current meters and foraminiferal tests

Abstract

Between June 1989 and July 1990 a mooring equipped with four current meters and one sediment trap was deployed in the Fram Strait (79°N). From the material collected by the sediment trap two species of planktic foraminifera (Turborotalita quinqueloba and Neogloboquadrina pachyderma) were used to estimate the catchment area of the sediment trap. For each species and size class the individual sinking velocity was used to determine the residence time in the water column. The residence time is the major factor that determines the drift of the tests. Using current-meter data trajectories were calculated to estimate the catchment area and the area of sedimentation. The maximum distance the smallest tests have been transported, while sinking from the surface $\text{(∼100}\text{m}\text{)}$ to the sediment trap (1125 m), is approximately 275 km, with maximum trajectory lengths of up to 300 km. Depending mainly on the grain size of the tests, the distances and trajectories have means lying between 25 and 50 km for N. pachyderma and 50 and 100 km for T. quinqueloba, respectively. Considering the whole sampling period, the estimated catchment areas are $\text{23,900}\text{km}{}^2$ for all N. pachyderma tests and $\text{33,300}\text{km}{}^2$ for all T. quinqueloba tests, respectively. The respective areas of sedimentation are estimated to range from $\text{3400}\text{km}{}^2$ to $\text{3500}\text{km}{}^2$.

Introduction

During the recent decades the particle flux in the world's oceans has received particular interest. Investigations on particle flux are available from nearly all oceans as well as from larger fresh water lakes (e.g. Ittekkot et al., 1996). During their gravity-induced vertical descent through the water column the particles, isolated or in aggegates, are subject to horizontal transport induced by currents. Thus the point where a particle starts the sedimentation process and the point where it finally will be deposited at the sea floor (or collected in a sediment trap) are normally at different locations. For particle-flux analyses using sediment traps and for palaeoenvironmental reconstructions using sediment cores, both aiming to describe or to be related to surface-ocean conditions, it is of particular interest to know over which surface-ocean area a respective sample integrates.

The aim of this paper is to estimate the catchment area for foraminifera of a sediment trap. For this purpose we combine data from current meters and a sediment trap moored in the Fram Strait. Other investigations used theoretical models (Deuser et al., 1988; Siegel et al., 1990) to determine the catchment area of a sediment trap. In this study, we use the tests of two species of planktic foraminifera of various size classes collected in the sediment trap as tracers and calculate their three-dimensional trajectories with respect to current meter data and sinking speed.

The Fram Strait is the deepest connection between the northern Atlantic Ocean and the Arctic Ocean and therefore plays a major role for the exchange of deep water between these two regions. The near-surface current system in the Fram Strait consists of the West Spitsbergen Current (WSC), which transports water of Atlantic origin northward into the Arctic Ocean, and the East Greenland Current (EGC), which transports Polar water southward to lower latitudes. Both currents are topographically steered. South of Spitsbergen the East Spitsbergen Current (ESC) enters the Fram Strait from the Barent Sea, also advecting water of Polar origin (Fig. 1) (e.g. Aagaard et al., 1985; Johanessen, 1986; Quadfasel et al., 1987; Morison, 1991). The temperature distribution of a transect across the Fram Strait at 79°N for the summer season is shown in Fig. 2.

The mooring 0162 was positioned in the WSC off Spitsbergen at 78°52.58′N and 6°40.50′E (Fig. 1). Four Aanderaa current meters were spaced vertically at a distance of approximately 500 m apart. A HDW SMT 230 sediment trap with $\text{0.5}\text{m}{}^2$ catchment area and 20 sample cups was placed at 1125 m water depth (for details refer to Table 1). Bottom depth was 1676 m. The mooring remained in its position at the continental slope between June 4, 1989 and July 24, 1990. The sediment trap opened on June 5, 1989 and took 11 samples of 16.5 days, 4 samples of 33 days and 5 samples of 16.5 days successively. The sampling period ended on July 6, 1990 (see Table 2). A sketch of the mooring structure is superimposed on the temperature distibution plot in Fig. 2.

Since this paper focuses on the catchment area of the sediment trap, we give only a brief overview on the flux pattern of planktic foraminifera at this site. The foraminifera in the sediment trap samples were picked individually under wet solution. Here, only foraminifera were selected that had sunk as isolated particles and not in aggregates. During this procedure the tests were counted, their size was measured in 16-μm steps (equal to one unit on the binocular scale). The specimens were determined to species level and examined for plasma content. For a more detailed description of the handling of the samples and results on the foraminiferal data see Carstens and Wefer (2000).

The fauna was dominated by Neogloboquadrina pachyderma (25%) and Turborotalita quinqueloba (57%). These species are observed nearly throughout the year. The flux pattern is given in detail in Carstens and Wefer (2000). Both species have their maximum flux in summer, from July to August. During winter time only single individuals contribute to the annual flux. Nearly all tests found in the samples were empty; about 2.3% for N. pachyderma and 1.5% for T. quinqueloba still contained plasma. From the size measurements, size-frequency distributions were calculated. The size distributions for both species vary during the sampling time. For an overview we give the size-frequency distributions for both species in Fig. 3 for all specimens throughout the year. N. pachyderma has larger tests (mean: $\text{193}\text{μm}$) than T. quinqueloba (mean: $\text{134}\text{μm}$). However, the latter species posesses lower sinking speeds (Table 3). The variations in the size during the year are shown in Fig. 4. In general, the foraminifera show larger tests during times of high foraminiferal flux for both species.

The four Aanderaa current meters registered velocity and direction of currents every hour. A filter was used on the current meter data to suppress the tidal signal. In Fig. 5a we show the mean current velocities for each sampling interval. In general, near-surface current meters show higher velocities than the deeper ones. Note the lower speeds during intervals 7–10. From sampling interval 12–19 an enhanced change in the direction and higher current velocities are observed, the latter especially at the deeper meters. As an example, detailed current meter data for two selected intervals is shown in Fig. 5b.

The catchment area was estimated by constructing three-dimensional trajectories. The trajectories display the path the tests traversed from the sea surface (i.e. level of 100 m depth) to the deep water (Fig. 6). With the sediment trap serving as the starting point a new position of the test with respect to the vertical axis is calculated with a time step of half an hour. Thus, every 30 min the velocity components with respect to space and time are determined by an interpolation between the nearest two current meters and a new position can be estimated in the next horizontal plane. This principle is persued until the level of the uppermost or lowermost current meter is reached. The time step of half an hour was selected to obtain a minimum spatial resolution of 10 m for the highest sinking velocities. For this reason the spatial resolution increases for smaller tests.

Corresponding to the trap sampling intervals, the current-meter data were divided into 20 sections. For the investigation the size frequencies of the foraminifera N. pachyderma and T. quinqueloba were used. Each size class from both species was taken into account because of the intrinsic sinking velocities. The sinking speeds used in this study are given in Table 2. Each of the 20 sample cups represents a defined time interval during which the velocity data show fluctuations in magnitude and direction. To cope with this difficulty one trajectory was calculated every 24 h during the opening interval to reach a sufficient temporal resolution.

With this method a large number of trajectories are determined describing the paths that the tests traversed or could have traversed if they had not been hindered by the trap. Each trajectory has a defined position when it reaches the surface or the bottom level. The end points of the trajectories of all species, size classes and intervals define the catchment and the sedimentation area.